31. RESULTS OF DOWNHOLE GEOPHYSICAL LOGGINGHOLE 396B, DSDP LEG 46

R. James Kirkpatrick1, Scripps Institution of Oceanography, University of California at San Diego, La Jolla, California

INTRODUCTION

The downhole logging of Hole 396B on Deep Sea DrillingProject Leg 46 was the first attempt to obtain in situ physicalproperty data in Layer 2 of the oceanic crust, and one of thefew published logging attempts in basaltic rocks (Zablocki etal., 1974; Keller et aL, 1974; Siems et al., 1973, 1974;Crosby and Anderson, 1971). Hole 396B is about 160 kmeast of the Mid-Atlantic Ridge at 22°59.14'N, 43°30.90'W.The oldest sediments cored from above the basalt are middleMiocene, about 13 million years old (Leg 46 ShipboardParty, this volume). The depth of the hole is 405.5 meters.The hole was logged from the base of the casing at 163 metersto about 370 meters, depending on the tool. The top of thebasement is at 150.5 meters (Leg 46 Shipboard Party, thisvolume).

The original objectives of the logging program were todetermine, in situ, physical properties and the positions of thelithologic contacts in the hole, especially in intervals of poorrecovery. Within the limitations of the tools used, theseobjectives were met.

Because logging instruments are not standardoceanographic tools, this paper will first present a summaryof the tools used and their limitations. It will then present thedata and address the following questions:

1) What are the tools measuring?2) How precise are the measurements?3) How representative of the in situ values are the data?4) How well can the logs be used to distinguish different

rock types?5) How well can the logs be used to determine lithologic

boundaries?6) What do the logging data tell us about the state of the

upper part of Layer 2 at this site?

LOGGING PROCEDURE AND TOOLS

The logging tools were provided by Schlumberger, Ltd.,and include compensated sonic velocity, compensateddensity,(gamma-gamma) compensated neutron porosity,electrical resistivity (Dual Induction - Laterolog 8), andnatural gamma ray. No caliper data were obtained.Temperature data are reported by Erickson and Hyndman(this volume). Because of the drill string's minimum 3% inch(9.84 cm) inside diameter, none of the positioning devices(centralizers or decentralizers) on any of the tools could be

'Present address: Department of Geology, University of Illinois, Ur-bana, Illinois.

used. This lack of positioning devices is an importantrestriction on quantitative use of the data, because thepositions of the tools (mostly about 3% in. [8.6 cm] O.D.) inthe hole (nominally 10 in. [25.4 cm]) are not well known.The tools are mostly of the compensated type (more than onereceiver), however, and the effects of mispositioning do notseem to be great if average data are used. Because of thisproblem, only average data for intervals 4 meters or longerwill be used.

The following discussion of the tools comes primarilyfrom the Schlumberger Log Interpretation manual(Schlumberger, 1972a, b).

The sonic velocity tool (compressional velocity only)consists of two sets of one transmitter and two receivers each.The interval transit time is obtained by alternately pulsingeach transmitter and averaging the travel time differencebetween the two receivers for each set. This compensationprocedure, similar to reversing a seismic refraction line,eliminates most of the sonde-tilt and hole-size effects. Thedata are recorded continuously as travel time (µ,s/ft), and nocorrections to the data on the log are necessary. Because nopositioning devices were used on any of the tools, these dataare probably the most accurate of all the logging data (R.Aguilar, personal communication). The tool investigatesonly a few centimeters into the rock.

The density tool used is also compensated. It consists of adirectional gamma ray source and two directional gamma raydetectors at different distances from the source. Thedifference in count rate between the two detectors isconverted to electron density in the rock, which is convertedto mass density. For most rock types, including basalts, nocorrections to the log data are necessary. The toolinvestigates to a depth of about 15 cm. The mud-cakecorrection, which has already been made, is plotted on thelog along with the density.

The neutron tool used is compensated in a way similar tothe density log. It consists of an Am-Be neutron source andtwo neutron detectors at different distances from the source.Comparison of the response of the two detectors eliminatesmuch of the sonde-tilt and hole-size effects. The neutronsrespond primarily to hydrogen atoms, although boron andchlorine have some effect. Since most but not all of thehydrogen in the rock is in pore water, it is necessary to makesome corrections which are not considered automatically inthe data reduction process. For Hole 396B some of theseeffects are significant. Because the temperature in the holewas low, but the pressure high (about half a kilobar), thereis a pressure correction of minus 10 per cent of the log value(C. Clavier, personal communication). In addition, there is

401

R. J. KIRKPATRICK

a minus 4 porosity per cent hole-size correction (10 in. holeversus an 8 in. standard condition) and a minus 2 porosityper cent bound-water correction. If the tool was not at theside of the hole, there is also a minus 3 porosity per cent perinch standoff correction. If the tool was in the center of thehole there is thus a minus 11 porosity per cent correction. InHole 396B the tool was most probably lying on the side ofthe hole, even without the decentralizers, because the holewas almost certainly not vertical (R. Aguilar, personalcommunication). Therefore, no standoff correction has beenmade. The depth of investigation of the neutron tooldepends on water content, but is less than about a foot.

The resistivity tool used (Dual Induction-Laterolog 8)consists of two induction tools which investigate to about 60inches (ILD) and 40 inches (ILM), and a focused electrodetool (LL8) which investigates closer to the hole. The twoinduction tools measure conductivity (which is converted toresistivity) in a horizontal plane perpendicular to the hole.The electrode tool measures components of both thehorizontal and vertical resistivity. Because the current iscarried primarily by the pore fluid, it is necessary to know theresistivity of the pore fluid to make quantitativeinterpretations of the resistivity data. The downholetemperature measurements of Erickson and Hyndman (Leg46 Shipboard Party, this volume) indicate that this fluid isseawater at about 2.5°C. This will be assumed in this paper.

The natural gamma ray tool used measures the totalgamma ray activity reaching the sonde. In most rocks thisactivity may be ascribed to 40K and the elements of theuranium and thorium decay series. Quantitativeinterpretation of the data is difficult, and will not beattempted here. Intervals of high natural gamma ray activitydo, however, correspond to intervals of low sonic velocityand density and high porosity. The higher activities almostcertainly arise from the potassium content of palagonitebreccia, which is higher than in unaltered basalt.

DATA

The five logs obtained are presented, along with ageneralized lithologic column, in Figure 1. Four majorlithologic units were identified in the basement (Leg 46Shipboard Party, this volume). The interval from 150.5meters to 235.5 meters (Units 1 and 2) is a sequence ofsparsely phyric basalt containing olivine and plagioclasephenocrysts, with some palagonite breccia and lithifiednannofossil ooze in the upper part. The section between235.5 meters and 243 meters (Unit 3) is a sparsely phyricbasalt flow. From 243.0 to 310 meters (Unit 4) is an olivineand plagioclase phyric basalt pillow sequence. The intervalfrom 310.0 to 405.5 meters (Units 5 through 8) is a sequenceof basaltic pillows, breccia, sand, and gravel; recovery in thisinterval was very poor.

The average values for the downhole data to be usedquantitatively, and coring and recovery data and a lithologiccolumn, are presented in Figure 2. Where the values aremarked O. S., the data are off scale on the processed logs, forwhich no backup scale was provided. The values, ifpresented, are taken from the on-site optical logs. Bothinterval travel time (µs/ft) and sonic velocity (km/s) arepresented for the sonic log. For the neutron log both the

uncorrected and corrected values are given. Data for all threeresistivity instruments are presented, although the ILD andLL8 fall so close to each other that they are plotted together.

The resistivity data require some special explanation. Theobservation is that the deep-investigating ILD and theshallow-investigating LL8 always track close to each other,and the intermediate-investigating ILM tracks with the othertwo in low-resistivity intervals and below in high-resistivityintervals (good basalt). This is because of the large boreholecorrection that must be applied to the ILM log at highresistivities (C. Clavier, personal communication). Thecorrection is necessary because a significant amount of theinduced current sensed is in the borehole. The corrections tothe ILD and LL8 logs are small, about 10 per cent. Given thelack of positioning devices, this correction is probably lessthan the accuracy of the data, and will not be made.

DISCUSSION

Before the logging data can be used to make geological orgeophysical interpretations, it is important to understandwhere the physical properties are being measured, what theprecision of the data is, and how representative the data are ofthe in situ conditions.

It is clear from Figures 1 and 2 that the measurements areof rocks and that none of the values are for water, eventhough the hole may be washed out, as is likely in the sand.From the available descriptions of the tools, it seems thatexcept for the induction tools the measurements are comingfrom within about a foot of the borehole.

Plots of the different data against each other can be used totest the precision of the data. For the Leg 46 data they seem toindicate that the precision is comparable to that of theshipboard laboratory data.

Probably the best test is the porosity-density plot shown inFigure 3. The position of a data point on this plot is related tograin density, which should be about 2.8 to 2.9 g/cm3 forthese rocks (Leg 46 Scientific Party, this volume; Hyndmanand Drury, 1976; Christensen and Salisbury, 1975). Thevalues for most intervals in Hole 396B fall within this range.In addition, in the high density range the log and laboratorydata overlap. The point labeled sand is from the lowestdensity part of the inferred sand and gravel interval, andprobably reflects inaccurate measurements caused bycollapse of the hole. The hole terminated because the drill bitstuck in this material. The palagonite breccia, labeled " P , "should have lower grain densities. That this does not appearto be true for the logs, as the one shipboard palagonitemeasurement indicates, may be a result of an insufficientbound-water correction for the palagonite.

It appears that for most of the data points the precision ofmeasurement, even without the positioning devices, is a fewporosity per cent and 0.1 or 0.2 g/cm3 for density, if the twotools are inaccurate, they are inaccurate in proportion to eachother and in the same direction.

Velocity and density (Figure 4) also show a consistentrelationship. Most of the logging data fall along the linear fitfor oceanic basalts of Christensen and Salisbury (1975). Thisline is also close to a linear extrapolation of the shipboarddata for Hole 396B (Leg 46 Shipboard Party, this volume).Again, in the high density range the laboratory and log data

402

DOWNHOLE GEOPHYSICAL LOGGING

170-

190-

210-

230-

250-

8 270-

290-Q.

Q

310-

330-

350-

370-

LJJ O

oocCO CO

EL ca

HI α

> h-H -io. <GC CO

O<CL. 00

δ<< o_ l LU< CCQ. CO

Q

< LU

INTERVALGAMMA RAY DENSITY POROSITY TRANSIT TIME(API UNITS) (gm/cmJ) INDEX (%) (µ sec/ft)

0 70 2.0 2.5 3.0 60 30 0 140 90 40

1

: :::

:==:::= S-,

i I 5 5 • i i:s«

: : : : : : : |

m::E:::±:

ZIZ_

EÈ=

::::::: +

i

r_:_ + . . .

• " = _ ? • 2 L •~

: [it

-- : = ° ¥~~

TTi- i H- t 5:=Φ:

1 ÈĖIĖĖS

=--f:r;:;:-L H._.

—r~j~rt—r —~~

5'r. : } • : : : :

m l } : I::: rj

==^pr: :: :r:

" " I T T i : : :

Up— ~ T ~ t r • • • •

i It: : : : :- -ü: : : : :

f::: ::::

1___

=•-== = i —

--V----3

• = = :;-+-

;===iiii5

:=: E E : 5 ;Ej

; : : = : ; ; ; ^= = i = j g ü ^

; = = - - - = - -

. . . Z\

JrrF 35

:: : : : : : : :

i=iEEEE±

: : : : : : : : |

: —? =-:|

::::::;::

WrWt 1

- : - - - : : ; |

:::s = = = = S

:; = . _ : E J E _ :

1= = - Φ • 8 — —

^aEEr . . .r

ET|FTiTF

-TfftFFF

:: : m

" 3 " ' ~ t 1 i l l

π iK, 3E E S ^

''̂ ~f-f~f ; : r

i t r: : : : : -T | | ^ ; : : : ._

4 = = t rr1111|1111 j

RESISTIVITY(ohm-m)

1.0 10 1001000

Figure 1. Processed logs for Hole 396B, with generalized lithologic column.

403

R. J. KIRKPATRICK

160 r

CORE NUMBERand RECOVERY

INTERVALTRANSIT TIME SOUND VELOCITY DENSITY

(µ sec/ft) (km/sec) (gm/cm3)140 90 40 1 3 5 2.0 2.5 3.0

. i

NEUTRON POROSITY RESISTIVITY(ohm-m)

10

200

240

5_

6

7

8

9

10

11-12-

• 13

14

- 15

280

320

360

400

17

18]

19

20

J- 21

22

- 23 I

24

_ 25

26

27

28

29

- 30

31"

32

- 33

H •

• • M

O<o o<o<ocD- CQ

SOO LU_J CC

O. LU

L U I -co -Jcc <<wQ. <CO CQ

FLOW

£<CC CO

o<

<

<d

LU LUCC LL

< >

w<

off scale

r

LOG

I

\

i s

Figure 2. Interval averaged values for the Hole 396B log data, with Uthologic column and recovery (dark) for each core. O.S.indicates that the values were off scale on the processed logs, for which no backup scale was provided. Values come fromon-site optical logs.

overlap. At low densities the data fall well below thenonlinear fit of Christensen and Salisbury (1975). This mayreflect the unconsolidated nature of the low-density material.

The velocity and porosity data (Figure 5) from the logs alsofall along a linear extrapolation of the shipboard data. Again,where the shipboard and log data are in the same porosity

range, they are in agreement. The Wyllie relationship

tr- t,

where <f> is porosity, tiog is the log travel time, tg is the grain

404

DOWNHOLE GEOPHYSICAL LOGGING

• 396 B Logging Data

• 396B Laboratory DataMost Hole 396BLaboratory Data

6.0

2.0 2.5

Density (gm/cm^)

3.0

Figure 3. Corrected neutron porosity versus density for theHole 396B logs.

travel time, and tf is the fluid travel time (Wylie et al.,1956, 1958), falls below the observed values, except for thepalagonite breccia, at porosities above about 35 per cent. Oneof the palagonite breccias plots above the general trend,possibly because of an insufficient bound-water correction tothe porosity. The sand falls well below the general trend,perhaps because it is much less consolidated or because ofpoor data quality in this possibly collapsed part of the hole.

Figure 6 is a plot of formation resistivity factor versusporosity for the laboratory data (Christensen and Hyndman,this volume) and logging data for Hole 396B. The formationresistivity factor (F), the ratio of rock resistivity to pore waterresisitivity, must be used in any comparison of data, becausethe water restivity can vary. Christensen and Hyndmanobtained their data at room temperature with seawater (Rw =0.2 ohm-m) in the pores. The pore fluid in the hole is alsomost likely sea water, but at about 2.5°C (Erickson andHyndman, this volume). A fluid resistivity of 0.4 ohm-meterhas been assumed.

As with the other data, the logging data fall on a linearextension of the laboratory data, and the two types of datanearly overlap.

Formation factor has been found (Archie, 1942) to berelated to porosity, Φ, by the relationship

F = CΦ~m

5.0

£ 4.0

3.0

2.0

Hole 396B Laboratory Data

Hole 396B Logging Data

Most Hole 396B .Laboratory Data

Linear f i t toOceanic Basalt Data(Christensen and Salisbury, 1975)

1.5 2.0 2.5Density (gm/cm^)

3.0

Figure 4. Sonic velocity versus density for the Hole 396Blogs and shipboard sample data, along with the linear fitfor oceanic basalts of Christensen and Salisbury (1975).

where C is a constant and m is related to the way the pores aredistributed (Hyndman and Drury, 1976). If the pores arerandomly distributed, as in a normally cemented sandstone,m is about 2. For uncemented materials m is less than 2, andfor materials with less than randomly connected pores m isgreater than 2.

A linear least-squares fit of the log and laboratory datagives an exponent of 1.33. For the log data alone, the slope is1.80 and for the laboratory data alone, 1.2. This is in goodagreement with log and laboratory data for Hawaiian basalts,which have an exponent of about 0.9 to 1.0 (G. Keller,personal communication). It does not, however, agree withthe exponent of about 2.5 found by Hyndman and Drury(1976) for laboratory samples of DSDP Leg 37 basalts. Theorigin of this difference is not clear, but may be relatedto the clay-mineral distribution (M. Drury, personalcommunication).

For the log data, at least, this low exponent seems to implythat the pores are connected better than in a consolidatedsandstone, and that given the high porosities, thepermeability is probably high.

A most significant and intractable question is whether thedownhole data are representative of the in situ values awayfrom the hole. As noted above, the rock near the hole could

405

R. J. KIRKPATRICK

• Hole 396B Laboratory Data

• Hole 396B Logging Data

4.0

3.0

2.0

1.0

Sand

10 20 30Porosity (%)

40 50

Figure 5. Sonic velocity versus porosity for Hole 396B logand shipboard sample data.

1000

100

10

• 396B Logging Data

• 396B Laboratory Data(Christensen andHyndman, thisvolume)

0.01 0.1Porosity

0.5

Figure 6. Formation factor versus porosity (rock resistivity Ipore fluid resistivity) for Hole 396B log data, assuming apore fluid resistivity of 0.4 ohm-meters, along with lab-oratory measurements of Hole 39 6B samples, assuming afluid resistivity of 0.2 ohm-meters (Christensen andothers, this volume).

be fractured and the hole wall could be extremely irregular,giving rise to logging data which may not be correct for thecrust as a whole. Two different lines of evidence indicate thatthe log data probably represent the true values.

Comparison of the integrated sonic velocity from the logwith velocities for the upper part of Layer 2 determined bymarine refraction methods indicates that the downhole datalie within the observed range. The integrated velocity for theinterval logged, 163 to 370 meters, is 3.1 km/s. If the sandand gravel unit is excluded, and of all the DSDP holes whichhave penetrated 200 meters or more into Layer 2, this is theonly time such a thick unit has been found, the averagevelocity is 3.6 km/s. Both these values are much less than theaverage velocity of about 5.5 km/s determined for oceanfloor basalts by Christensen and Salisbury (1975) and for thereturned material at Site 396 (Leg 46 Shipboard Party, thisvolume). These are, however, within the range of 3.0 to 4.5km/s for upper Layer 2 velocities in 13 million-year-old crust(Houtz and Ewing, 1976). The average laboratory value forbasalt does not fall within their range until the crust is about20 million years old. At the very least, then, the log velocitiesare a lower boundary for the true velocity at this location, andare probably closer to the true velocity than those indicatedby the laboratory measurements.

Corroborating evidence comes from the variation ofresistivity with distance from the hole. If there was extensivedamage to the rock near the hole, we should expect the LL8resistivity to be lower than the ILD resistivity. In fact, theyare always very close to each other. This implies that, to theextent that formation damage would decrease the resistivity,there is little damage or the damage is constant out to at least60 inches from the hole (unlikely).

These two pieces of evidence, taken together, indicatethat, on the average, hole damage is not extensive and the logdata are probably not far from the true in situ values.

USE OF THE LOGGING DATA

One of the important uses of the downhole data is inidentifying rock types and lithologic boundaries in intervalswhere recovery is poor. A major objective of this firstexperiment with logging in the oceanic crust was todetermine how well logging can perform this task.

Comparison of the logs and the recovered core showsclearly that boundaries between grossly different rock typescan be distinguished well and that in many cases the rocktypes can be distinguished uniquely. All the importantboundaries can be picked on all the logs except the resistivitylog, and rock types can be distinguished using a combinationof tools. It should be noted that only lithologic changes whichsignificantly affect the physical properties can be detected.Changes in basalt chemistry or texture apparently cannot bedetected with the present set of tools.

Three major rock types were encountered in Hole 396B:basalt, palagonite (hydrated basaltic glass) breccia, andbasaltic sand and gravel (perhaps slightly cemented) (Leg 46Shipboard Party, this volume). All of these can bedistinguished one from another using the logging data. Thebasalt (210 to 240 and 256 to 315 m) can be distinguished byits relatively high sonic velocity (4.0 to 4.8 km/s), density(2.4 to 2.6 g/cm3), and resistivity (up to 95 ohm-m), andrelatively low porosity (13 to 21%). For the most part, thegood basalt falls along uniform trends on the

406

DOWNHOLE GEOPHYSICAL LOGGING

porosity-density, porosity-velocity, density-velocity, andformation factor-porosity plots.

The palagonite breccia (310 to 340 m and perhaps 163 to167 m) can be distinguished by its high velocity-porosityratio (Figure 5) and its high m value on the formationfactor-porosity plot (Figure 6). Both of these characteristicsmay be ascribable to an insufficient bound-water correctionto the neutron porosity data, resulting in an overestimate ofthe porosity. In principle, the grain density for the palagoniteshould be less than for basalt. That this is not the case inFigure 3 may again be the result of an overestimate of theporosity.

The sand and gravel unit (below 340 m) can easily bedistinguished by its low velocity (1.8 km/s), low density(1.55 g/cm3), and low resistivity (less than 10 ohm-m).

There are, however, intervals where correlation betweenthe logs and the core is not clear. This is especially true in theupper parts of the sparsely phyric and porphyritic pillowbasalt units. In the sparsely phyric basalt unit we recovered alittle palagonite breccia with pillow basalt to a depth of about174 meters. The high natural gamma ray activity also endshere. The highly irregular character of the logs in thisinterval, however, extends to a depth of about 210 meters.This discrepancy probably results not from lack of recoveryof breccia between 174 and 210 meters, but from openfractures or cavities (pillow centers?) which extend to 210meters. In the upper part of the porphyritic basalt unit, thevelocity, density, and resistivity are low and the porosityhigh, but only pillow basalt was recovered. Once again,palagonite breccia was not recovered and the natural gammaresponse is still low. It seems likely that there are openfractures or cavities here also. If there are, then pillowsequences may be identifiable by a change from highvelocity, density, and resistivity and low porosity at thebottom to lower velocity, density, and resistivity and higherporosity at the top, but with the data plotting on the generalbasalt trend on the porosity-density, porosity-velocity,density-velocity, and formation factor-porosity diagrams.

All the lithologic boundaries in Hole 396B can beidentified using any of the logs except resistivity, whichidentifies some. The aphyric pillow-basalt/basalt-flowcontact, however, cannot be identified at all, except perhapsby a slight porosity decrease in the flow. Theflow/porphyritic-basalt contact is identifiable in all logs, andthe porphyritic-basalt/palagonite-breccia contact isidentifiable except with resistivity. Position of thebreccia/sand-and-gravel contact is defined by the logs. Itseems likely that similar contacts could be determined inother holes if logging were available.

CONCLUSIONS — THE STATE OF THE CRUSTIf we can rely on the inferences drawn from the lack of

radial variation in resistivity and from comparison of theintegrated sonic velocity from the logs with the geophysicallydetermined velocities for the upper part of Layer 2, the datapresented in Figure 2 may be a good first approximation ofthe true physical state of the interval logged at Site 396.

Perhaps the most impressive result is the overall highporosity and low density throughout the hole. Even in themassive basalt flow the porosity appears to be about 13 percent — much higher than the average laboratory value ofabout 4 per cent. This seems to imply that the porosity is on a

scale larger than the core samples, i.e., large open fracturesand cavities. This is in agreement with the inferred conditionsat Mid-Atlantic Ridge Sites 332 and 333 (Hyndman andDrury, 1976). It is also in good agreement with theapparently good connection between the pores, indicated bythe porosity-formation factor plot. Such large openings,combined with the overall high porosity, would seem toimply high permeability through the sequence drilled.

One important conclusion from this first loggingexperiment is that more logging in the oceanic crust isneeded. Logging is the only way presently available todetermine a suite of in situ physical properties, and is the onlyway to determine lithologic boundaries in intervals of poorrecovery. There is also a need for additional tools to helpresolve some of the ambiguities in the present data. Intervalvelocity would go a long way toward finally determining theextent of disturbance around the bore hole, and a continuoustemperature log, and perhaps downhole fluid sampling,would insure proper interpretation of the resistivity data.

ACKNOWLEDGMENTSI wish to thank R.E. Boyce, J. Severns, J. Heirtzler, R.

Hyndman, C. Clavier, M. Drury, and G. Keller for their manyuseful discussions and comments during the course of this work,and Rudy Aguilar for collecting the original data. R.E. Boyce and J.Severns reviewed the manuscript.

REFERENCES

Archie, G.E., 1942. The electrical resistivity log as an aid indetermining some reservoir characteristics, Pet. Tech., v. 5.

Christensen, N.I. and Salisbury, M.H., 1975. Structure andconstitution of the lower oceanic crust, Rev. Geophys. SpacePhys., v. 13, p. 57-86.

Crosby, J.W. and Anderson, J.V., 1971. Some applications ofgeophysical well logging to basalt hydrogeology, GroundWater, v. 9, p. 13-20.

Houtz, R. and Ewing, J., 1976. Upper crustal structure as a functionof plate age, J. Geophys. Res., v. 81, p. 2490-2498.

Hyndman, R.D. and Drury, M., 1976. The physical properties ofoceanic basement rocks from deep-drilling on the Mid-AtlanticRidge, J. Geophys. Res., v. 81, p. 4042-4052.

Keller, G.V., Murray, J.C., Skokan, J.J., andSkokan, C.J., 1974.CSM research drill hole at summit of Kilauea Volcano, TheMines Mag., May, p. 14-18.

Schlumberger, Ltd., 1972a. Log Interpreation, Volume 1 -Principals, p. 113.

, 1972b. Log Interpretation, Volume 2 - Charts, p. 92.Siems, B.A., Crosby, J.W., Anderson, J.V., Bush, J.H., and

Webere, T.L., 1973. Final report, geophysical investigation ofWashington's ground water resources, Wash. State Univ. Col.of Engineering Res. Div. Rept. No. 73/15-81.

Siems, B.A., Bush, J.H., and Crosby, J.W., 1974. TiO2 andgeophysical logging criteria for Yakama basalt correlation,Columbia Plateau, Geol. Soc.Am.Bull., v. 85, p. 1061-1068.

Wyllie, M.R.J., Gregory, A.R., and Gardner, G.H.F., 1956.Elastic wave velocities in heterogeneous and porous media,Geophysics, v. 21, p. 41.

, 1958. An experimental investigation of factorsaffecting elastic wave velocities in porous media, Geophysics,v. 23, p. 459-493.

Zablocki, C.J., Tilling, R.I., Peterson, D.W., Christianson, R.I.,Keller, G.V., and Murray, J.C., 1974. A deep research drillhole at the summit of an active volcano, Kilauea, Hawaii,Geophys. Res. Letters, v. 1, p. 323-326.

407